Lancet Digital Health最新文献

{"title":"How can artificial intelligence transform the training of medical students and physicians?","authors":"Yilin Ning PhD ,&nbsp;Jasmine Chiat Ling Ong PharmD ,&nbsp;Haoran Cheng MPH ,&nbsp;Haibo Wang MPH ,&nbsp;Daniel Shu Wei Ting MD PhD ,&nbsp;Yih Chung Tham PhD ,&nbsp;Prof Tien Yin Wong MD PhD ,&nbsp;Nan Liu PhD","doi":"10.1016/j.landig.2025.100900","DOIUrl":"10.1016/j.landig.2025.100900","url":null,"abstract":"<div><div>Advances in artificial intelligence (AI), particularly generative AI, hold promise for transforming medical education and physician training in response to increasing health-care demands and shortages in the global health-care workforce. Meanwhile, challenges remain in the effective and equitable integration of AI technology into medical education and physician training worldwide. This Viewpoint explores the opportunities and challenges of such an integration. We study the evolving role of AI in medical education, its potential to enhance high-fidelity clinical training, and its contribution to research training using real-world examples. We also highlight ethical concerns, particularly the unclear boundaries of appropriate use of AI and call for clear guidelines to govern the integration of AI into medical education and physician training. Furthermore, this Viewpoint discusses practical constraints, including human, financial, and resource constraints, in AI integration, and emphasises the need for comprehensive cost evaluations and collaborative funding models to support the sustainable implementation of AI integration. A tight collaborative network between health-care institutions and systems, medical schools and universities, industry partners, and education and health-care regulatory agencies could lead to an AI-transformed medical education and physician training scheme that ultimately supports the adoption and integration of AI into clinical medicine and potentially brings about tangible improvements in global health-care delivery.</div></div>","PeriodicalId":48534,"journal":{"name":"Lancet Digital Health","volume":"7 10","pages":"Article 100900"},"PeriodicalIF":24.1,"publicationDate":"2025-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145233948","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Extension of the GRACE score for non-ST-elevation acute coronary syndrome: a development and validation study in ten countries","authors":"Florian A Wenzl MD ,&nbsp;Klaus F Kofoed DMSc ,&nbsp;Moa Simonsson MD ,&nbsp;Gareth Ambler PhD ,&nbsp;Niels M R van der Sangen MD ,&nbsp;Erik Lampa PhD ,&nbsp;Francesco Bruno MD ,&nbsp;Mark A de Belder MD ,&nbsp;Jiri Hlasensky PhD ,&nbsp;Matthias Mueller-Hennessen MD ,&nbsp;Maria A Smolle MD ,&nbsp;Peizhi Wang BMed ,&nbsp;José P S Henriques MD ,&nbsp;Wouter J Kikkert MD ,&nbsp;Henning Kelbæk DMSc ,&nbsp;Luboš Bouček MD ,&nbsp;Sergio Raposeiras-Roubín MD ,&nbsp;Emad Abu-Assi MD ,&nbsp;Jaouad Azzahhafi MD ,&nbsp;Matthijs A Velders PhD ,&nbsp;Thomas F Lüscher","doi":"10.1016/j.landig.2025.100907","DOIUrl":"10.1016/j.landig.2025.100907","url":null,"abstract":"&lt;div&gt;&lt;h3&gt;Background&lt;/h3&gt;&lt;div&gt;The Global Registry of Acute Coronary Events (GRACE) scoring system guides the management of patients with non-ST-elevation acute coronary syndrome (NSTE-ACS) according to current guidelines. However, broad validation of the sex-specific GRACE 3.0 in-hospital mortality model, and corresponding models for predicting long-term mortality and the personalised effect of early invasive management, are still needed.&lt;/div&gt;&lt;/div&gt;&lt;div&gt;&lt;h3&gt;Methods&lt;/h3&gt;&lt;div&gt;We used data of 609 063 patients with NSTE-ACS from ten countries between Jan 1, 2005, and June 24, 2024. A machine learning model for 1-year mortality was developed in 400 054 patients from England, Wales, and Northern Ireland. Both the in-hospital mortality model and the new 1-year mortality model were externally validated in patients from Sweden, Switzerland, Germany, Denmark, Spain, the Netherlands, and Czechia. A separate machine learning model to predict the individualised effect of early versus delayed invasive coronary angiography and revascularisation on a composite primary outcome of all-cause death, non-fatal recurrent myocardial infarction, hospital admission for refractory myocardial ischaemia, or hospital admission for heart failure at a median follow-up of 4·3 years was developed and externally validated in participants from geographically different sets of hospitals in the Danish VERDICT trial.&lt;/div&gt;&lt;/div&gt;&lt;div&gt;&lt;h3&gt;Findings&lt;/h3&gt;&lt;div&gt;The in-hospital mortality model (area under the receiver operating characteristic curve [AUC] 0·90, 95% CI 0·89–0·91) and the 1-year mortality model (time-dependent AUC 0·84, 95% CI 0·82–0·86) showed excellent discriminative abilities on external validation across all countries. Both models were well calibrated and decision curve analyses suggested favourable clinical utility. Compared with score version 2.0, both models provided improved discrimination and risk reclassification. The individualised treatment effect model effectively identified patients who would benefit from early invasive management on external validation. Patients with high predicted benefit had reduced risk of the composite outcome when randomly assigned to early invasive management (hazard ratio 0·60, 95% CI 0·41–0·88), whereas patients with no-to-moderate predicted benefit did not (1·06, 0·80–1·40; p&lt;sub&gt;interaction&lt;/sub&gt;=0·014). The individualised treatment effect model suggested that the group of patients with NSTE-ACS who benefit from early intervention might be incompletely captured by current treatment strategies.&lt;/div&gt;&lt;/div&gt;&lt;div&gt;&lt;h3&gt;Interpretation&lt;/h3&gt;&lt;div&gt;The updated GRACE 3.0 scoring system provides a validated, practical tool to support personalised risk assessment in patients with NSTE-ACS. Prediction of an individual’s long-term cardiovascular benefit from early invasive management could refine future trial design.&lt;/div&gt;&lt;/div&gt;&lt;div&gt;&lt;h3&gt;Funding&lt;/h3&gt;&lt;div&gt;Swiss Heart Foundation, University of Zurich Foundation, Kurt and Senta Herrmann Foundation, Theodor ","PeriodicalId":48534,"journal":{"name":"Lancet Digital Health","volume":"7 10","pages":"Article 100907"},"PeriodicalIF":24.1,"publicationDate":"2025-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145313905","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Characterising the design and methods of continuous glucose monitoring used in behavioural interventions to inform future research in prediabetes","authors":"Prof David S Black PhD ,&nbsp;Alaina P Vidmar MD ,&nbsp;Braden Barnett MD","doi":"10.1016/j.landig.2025.100904","DOIUrl":"10.1016/j.landig.2025.100904","url":null,"abstract":"<div><div>Digital health feedback technologies are expected to help address the projected 630 million individuals with prediabetes worldwide by 2045. This Viewpoint article characterises the historical use of continuous glucose monitoring (CGM) systems in behavioural research with a focus on the prediabetic population. We identified 19 peer-reviewed studies through a pragmatic literature review and reported key methodological features, including study design, sensor wear protocols, data masking strategies, the role of CGM in behavioural interventions, and approaches to generate CGM metrics. Based on our literature review, we propose four directions to advance CGM in behavioural intervention research in prediabetes: refining sampling strategies to focus recruitment on individuals with prediabetes to better understand metrics in this population; improving transparency in CGM feedback delivery protocols; reporting a comprehensive and targeted set of CGM metrics; and articulating principles that account for the effects of CGM use within behavioural interventions. This methodological characterisation of CGM is a starting point to enhance research quality and behavioural intervention effectiveness, particularly when integrating CGM systems aimed at supporting dietary, physical activity, or lifestyle modifications among people with prediabetes.</div></div>","PeriodicalId":48534,"journal":{"name":"Lancet Digital Health","volume":"7 10","pages":"Article 100904"},"PeriodicalIF":24.1,"publicationDate":"2025-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145446398","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Development and validation of a machine-learning model to reduce futile procurements in donations after circulatory death in liver transplantation in the USA: a multicentre study","authors":"Rintaro Yanagawa ,&nbsp;Kazuhiro Iwadoh PhD ,&nbsp;Toshihiro Nakayama MD ,&nbsp;Daniel J Firl MD ,&nbsp;Chase J Wehrle MD ,&nbsp;Yuki Bekki PhD ,&nbsp;Daiki Soma MD ,&nbsp;Jiro Kusakabe MD ,&nbsp;Yuzuru Sambommatsu MD ,&nbsp;Yutaka Endo PhD ,&nbsp;Kliment K Bozhilov MD ,&nbsp;Jenny H Pan MD ,&nbsp;Masaru Kubota PhD ,&nbsp;Koji Tomiyama PhD ,&nbsp;Masato Fujiki PhD ,&nbsp;Magdy Attia MD ,&nbsp;Prof Marc L Melcher PhD ,&nbsp;Prof Kazunari Sasaki MD","doi":"10.1016/j.landig.2025.100918","DOIUrl":"10.1016/j.landig.2025.100918","url":null,"abstract":"&lt;div&gt;&lt;h3&gt;Background&lt;/h3&gt;&lt;div&gt;The number of liver transplants from donors after the circulatory determination of death continues to increase, helping to alleviate the existing organ shortage. However, the rate of attempted but subsequently terminated procurements, known as futile procurements, remains high—mainly because many potential donors do not progress to death within a timeframe after extubation that maintains the suitability of the organ for donation. Futile procurements pose considerable financial and workload burdens to the transplant system. We aimed to develop and validate a machine-learning model to better predict progression to death and reduce futile procurements in cases of donation after circulatory death (DCD).&lt;/div&gt;&lt;/div&gt;&lt;div&gt;&lt;h3&gt;Methods&lt;/h3&gt;&lt;div&gt;This study included data from 2221 donors from six centres in the USA. Using a retrospective dataset obtained from 1616 donors between December 1, 2022, and June 30, 2023, we developed a prediction model using the Light Gradient Boosting Machine (LightGBM) framework, with neurological, biochemical, respiratory, and circulatory parameters as predictors. The model was validated retrospectively with data from 398 donors (July 1–Aug 31, 2023) and prospectively with data from 207 donors (March 1–Sept 30, 2024). The performance of the model was evaluated through the area under the receiver operating characteristic curve (AUC), accuracy, futile procurement rate, and missed opportunity rate. We also compared the performance of the model with that of two existing risk-prediction tools (the DCD-N score and the Colorado Calculator) and surgeon predictions.&lt;/div&gt;&lt;/div&gt;&lt;div&gt;&lt;h3&gt;Findings&lt;/h3&gt;&lt;div&gt;Of the 2221 DCD donors in this study, 1260 progressed to death, 927 of whom died within 30 min after extubation. Cross-validation of the LightGBM model yielded AUCs for predicting donor progression to death of 0·833 (95% CI 0·798–0·868) at 30 min, 0·801 (0·767–0·834) at 45 min, and 0·805 (0·770–0·841) at 60 min after extubation. This performance was maintained in both retrospective (0·834 [0·772–0·891], 0·819 [0·757–0·870], and 0·799 [0·737–0·855]) and prospective (0·831 [0·768–0·885], 0·812 [0·749–0·874], and 0·805 [0·740–0·868]) validation cohorts. Compared with surgeon predictions, the LightGBM model had lower futile procurement rates (0·195 &lt;em&gt;vs&lt;/em&gt; 0·078, respectively), higher accuracy in cases of poor intersurgeon agreement (0·08 &lt;em&gt;vs&lt;/em&gt; 0·29) at 30 min, and similar missed opportunity rates (0·155 &lt;em&gt;vs&lt;/em&gt; 0·167). By contrast, the DCD-N score had AUCs of 0·799 (95% CI 0·730–0·860) at 30 min, 0·760 (0·695–0·824) at 45 min, and 0·739 (0·668–0·801) at 60 min, and the Colorado Calculator had AUCs of 0·694 (0·616–0·768), 0·669 (0·596–0·742), and 0·663 (0·585–0·736) at the same timepoints.&lt;/div&gt;&lt;/div&gt;&lt;div&gt;&lt;h3&gt;Interpretation&lt;/h3&gt;&lt;div&gt;We show that, compared with surgeon predictions and existing risk-prediction tools, our machine-learning model can enhance the accuracy of the prediction of progression","PeriodicalId":48534,"journal":{"name":"Lancet Digital Health","volume":"7 10","pages":"Article 100918"},"PeriodicalIF":24.1,"publicationDate":"2025-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145524521","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Reducing futile donation after circulatory death procurement with machine learning","authors":"Bima J Hasjim ,&nbsp;Mamatha Bhat","doi":"10.1016/j.landig.2025.100932","DOIUrl":"10.1016/j.landig.2025.100932","url":null,"abstract":"","PeriodicalId":48534,"journal":{"name":"Lancet Digital Health","volume":"7 10","pages":"Article 100932"},"PeriodicalIF":24.1,"publicationDate":"2025-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145524601","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The Jevons Paradox in global health: efficiency, demand, and the AI dilemma","authors":"Michael JA Reid ,&nbsp;Bilal Mateen","doi":"10.1016/j.landig.2025.100928","DOIUrl":"10.1016/j.landig.2025.100928","url":null,"abstract":"","PeriodicalId":48534,"journal":{"name":"Lancet Digital Health","volume":"7 10","pages":"Article 100928"},"PeriodicalIF":24.1,"publicationDate":"2025-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145524536","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Lessons learned from the IBD-BOOST trial: an opportunity for the next generation of behavioural self-management research in inflammatory bowel disease","authors":"Sonya Kuzminski ,&nbsp;Laurie Keefer","doi":"10.1016/j.landig.2025.100929","DOIUrl":"10.1016/j.landig.2025.100929","url":null,"abstract":"","PeriodicalId":48534,"journal":{"name":"Lancet Digital Health","volume":"7 10","pages":"Article 100929"},"PeriodicalIF":24.1,"publicationDate":"2025-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145476750","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Artificial intelligence in women’s cancers: innovation and challenges in clinical translation","authors":"Prof Andrea G Rockall FRCR ,&nbsp;Selina MY Chiu MRCOG ,&nbsp;Eric O Aboagye PhD ,&nbsp;Magnus Dustler PhD ,&nbsp;Christina Fotopoulou MD ,&nbsp;Sadaf Ghaem-Maghami FRCOG ,&nbsp;Alexandra Taylor FRCR ,&nbsp;Sophia Zackrisson MD","doi":"10.1016/j.landig.2025.100940","DOIUrl":"10.1016/j.landig.2025.100940","url":null,"abstract":"<div><div>Artificial intelligence (AI) is set to transform the care of women with cancer. From early detection via digital phenotyping to diagnosis, treatment, and follow-up, innovative AI applications are rapidly emerging across the cancer care continuum. AI-assisted mammographic screening for breast cancer has been clinically translated, and AI-based contouring in radiotherapy is streamlining treatment planning and improving consistency of cancer care. Research areas include radiomic analysis for ovarian tumour characterisation, machine learning for endometrial cancer subtyping, and automated assessment for cervical cancer screening. Challenges such as data scarcity and tumour heterogeneity in gynaecological cancers hinder the development of robust AI models, a problem further compounded by the limited availability of large, prospective validation cohorts. Emerging generative AI and multimodal AI systems hold promise to address these limitations by leveraging large-scale, diverse training datasets. Building trust in AI systems will require rigorous prospective real-life validation, regulatory oversights, and well-defined legal frameworks. A key opportunity exists to develop inclusive, clinically meaningful AI devices across all women’s cancers, driven by rapid advances in AI in health care and strengthened by national and international initiatives promoting health-care innovation. Through multidisciplinary collaboration, AI has the potential to move beyond research and help in early diagnoses and provide personalised treatment strategies. In this Series paper, we review AI developments in breast and gynaecological cancers, including applications in clinical adoption and those actively being developed to address unmet needs in early detection, characterisation, treatment, and prognostication.</div></div>","PeriodicalId":48534,"journal":{"name":"Lancet Digital Health","volume":"7 10","pages":"Article 100940"},"PeriodicalIF":24.1,"publicationDate":"2025-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145530889","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Leveraging deep learning applied to chest radiograph images to identify individuals at high risk of chronic obstructive pulmonary disease: a retrospective model validation study","authors":"Saman Doroodgar Jorshery MD ,&nbsp;Jay Chandra BA ,&nbsp;Anika S Walia BA ,&nbsp;Audra Stumiolo MS ,&nbsp;Kristin Corey MD ,&nbsp;Seyedeh Maryam Zekavat MD ,&nbsp;Aniket N Zinzuwadia MD ,&nbsp;Krisha Patel ,&nbsp;Sarah Short MPH ,&nbsp;Jessica L Mega MD ,&nbsp;R Scooter Plowman MD ,&nbsp;Neha Pagidipati MD ,&nbsp;Prof Shannon S Sullivan MD ,&nbsp;Prof Kenneth W Mahaffey MD ,&nbsp;Prof Svati H Shah MD ,&nbsp;Prof Adrian F Hernandez MD ,&nbsp;Prof David Christiani MD ,&nbsp;Hugo J W L Aerts PhD ,&nbsp;Jakob Weiss MD ,&nbsp;Michael T Lu MD ,&nbsp;Vineet K Raghu PhD","doi":"10.1016/j.landig.2025.100903","DOIUrl":"10.1016/j.landig.2025.100903","url":null,"abstract":"<div><h3>Background</h3><div>Chronic obstructive pulmonary disease (COPD) is a leading cause of mortality, yet early detection remains challenging. This study assessed whether deep learning applied to routine outpatient chest radiographs (CXRs) can identify individuals at high risk of incident COPD.</div></div><div><h3>Methods</h3><div>Using cancer screening trial data, we previously developed a convolutional neural network (CXR-Lung-Risk) to predict lung-related mortality from a CXR image. In this retrospective model validation study, we externally validated whether CXR-Lung-Risk was associated with incident COPD from routine CXRs. We identified outpatients without lung cancer, COPD, or emphysema who had a CXR taken from Jan 1, 2013, to Dec 31, 2014, at Massachusetts General or Brigham and Women’s Hospitals in Boston, MA, USA. The primary outcome was 6-year incident COPD. Discrimination was assessed using area under the receiver operating characteristic curve (AUC) compared with the TargetCOPD clinical risk score. All analyses were stratified by smoking status. A secondary analysis was conducted in the Project Baseline Health Study (PBHS) to test associations between CXR-Lung-Risk with pulmonary function and plasma protein abundance. The PBHS study is registered with <span><span>ClinicalTrials.gov</span><svg><path></path></svg></span>, <span><span>NCT03154346</span><svg><path></path></svg></span>.</div></div><div><h3>Findings</h3><div>The primary analysis consisted of data from 12 550 ever-smokers (mean age 62·4 years [SD 6·8], 6135 [48·9%] male, 6415 [51·1%] female) and 15 298 never-smokers (mean age 63·0 years [8·1], 6550 [42·8%] male, 8748 [57·2%] female). 1562 (12·4%) of 12 550 ever-smokers and 580 (3·8%) of 15 298 never-smokers developed COPD within 6 years. CXR-Lung-Risk had additive predictive value beyond the TargetCOPD score for 6-year incident COPD in both ever-smokers (CXR-Lung-Risk + TargetCOPD AUC 0·73 [95% CI 0·72–0·74] <em>vs</em> TargetCOPD alone AUC 0·66 [0·65–0·68], p&lt;0·0001) and never-smokers (CXR-Lung-Risk + TargetCOPD AUC 0·70 [0·67–0·72] <em>vs</em> TargetCOPD alone AUC 0·60 [0·57–0·62], p&lt;0·0001). In secondary analyses of 2097 individuals in the PBHS, CXR-Lung-Risk was associated with worse pulmonary function and with abundance of SCGB3A2 (secretoglobin family 3A member 2) and LYZ (lysozyme), proteins involved in pulmonary physiology.</div></div><div><h3>Interpretation</h3><div>In this external validation, a deep-learning model applied to routine CXR images identified individuals at high risk of incident COPD, beyond known risk factors. Patients at high risk might benefit from diagnostic spirometry and subsequent preventive care.</div></div><div><h3>Funding</h3><div>Verily Life Sciences, San Francisco, California.</div></div>","PeriodicalId":48534,"journal":{"name":"Lancet Digital Health","volume":"7 9","pages":"Article 100903"},"PeriodicalIF":24.1,"publicationDate":"2025-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145366195","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Computer-aided reading of chest radiographs for paediatric tuberculosis: current status and future directions","authors":"Mackenzie DuPont MD ,&nbsp;Robert Castro MPH ,&nbsp;Sandra V Kik PhD ,&nbsp;Megan Palmer MBChB ,&nbsp;Prof James A Seddon PhD ,&nbsp;Devan Jaganath MD","doi":"10.1016/j.landig.2025.100884","DOIUrl":"10.1016/j.landig.2025.100884","url":null,"abstract":"<div><div>Computer-aided detection (CAD) systems for automated reading of chest x-rays (CXRs) have been developed and approved for tuberculosis triage in adults but not in children. However, CXR is frequently the only adjunctive tool for clinical assessment in the evaluation of paediatric tuberculosis in primary care settings, and children would benefit from CAD models that can detect their unique clinical and radiographic features. To advance CAD for childhood tuberculosis, large, diverse paediatric CXR datasets linked to standardised tuberculosis classifications are required. These datasets would be used to train and validate paediatric-specific models for tuberculosis screening, diagnosis, and severity stratification. Previous studies on CAD algorithms for reading paediatric CXRs have highlighted promising approaches, including the use of transfer learning with existing deep learning models. Including data from children in CAD models is essential to improve equity and reduce the global burden of tuberculosis disease.</div></div>","PeriodicalId":48534,"journal":{"name":"Lancet Digital Health","volume":"7 9","pages":"Article 100884"},"PeriodicalIF":24.1,"publicationDate":"2025-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144974477","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}